Research On Network Traffic Prediction Method Of Industrial Control System Based On SCSO-GRU Model

As an important component of national key infrastructure,industrial control system is very important to the success of industry.Power production and distribution,water supply and treatment rely on industrial control system to achieve automatic control.Industrial control network is the most critical part of the whole industrial control system,which determines whether the information of the whole industrial control system can be transmitted and operated normally and stably.With the continuous cross integration of industrialization and information technology,the industrial control system began to use general network equipment and general software and hardware gradually.As a result,the closure of ICs was broken.If the flow of industrial control network is abnormal for a long time,it will not only affect the real-time and certainty of the network,but also may cause serious security accidents.From the perspective of reliability,it is one of the basic requirements for the reliability of industrial control system to grasp the trend of network traffic through prediction.From the perspective of security,the traffic prediction model can be combined with the network anomaly detection system of industrial control system,to improve the possibility of avoiding malicious attacks.From the perspective of certainty,in some specific security control systems(such as nuclear power plant safety level I & C system),the relevant safety control behavior needs to guarantee the deterministic demand,and the protection action must be completed within a certain time when the system safety condition is extremelydeteriorated,so there is a high demand for the certainty of network communication.This paper comes from the project of theoretical deduction,simulation verification and real platform verification of a safety level DCS system design scheme from theoretical modeling,simulation and other aspects proposed by a domestic design institute.Based on the above analysis,this paper proposes a GRU neural network traffic prediction model(SCSO-GRU model)based on the optimization of SCSO algorithm to predict and analyze the industrial control network traffic.The main work and contributions are summarized as follows:(1)We completed the project requirements basically.We have completed the work of **** for a research and design institute,and carried out theoretical demonstration,simulation,actual platform test and other work for their designed network structure and protocol.Finally,we realized the verification of their design scheme.(2)Based on SCSO-GRU model,the modeling and prediction of industrial control network traffic.In this paper,through the modeling and prediction of two kinds of industrial control network traffic,it is found that the GRU model based on the optimization of SCSO algorithm has a very good effect,which is better than ARIMA model and the modeling and prediction ability of using GRU and LSTM directly,in terms of prediction accuracy and convergence speed.It provides an optional model for the prediction of industrial control network traffic.